メニュー
Format one read-only optimizer plan into a concise decision report.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
Apply one bundled semantic material preset through the supervised shader tuning channel.
Propose normalized outfit names from read-only binding evidence while unsafe object and parameter renames stay blocked.
Add concise, read-only review sections to the standard validation report.
Scan the current avatar and summarize safe, read-only findings.
SOC 職業分類に基づく
| name | optimizer-report-helper |
| title | Optimizer Report Helper |
| description | Format one read-only optimizer plan into a concise decision report. |
| permission-mode | preview |
| risk-level | medium |
| allowed-tools | ["vrcforge_optimization_plan"] |
| entrypoint-tool | vrcforge_optimization_plan |
| support-files | ["workflows/optimizer-report-helper.json"] |
| test-command | python -m pytest tests/test_example_skill_packages.py -q |
Use this skill when a user wants optimizer scan results explained without
applying optimization. Invoke vrcforge_optimization_plan exactly once, then
format its returned plan into Summary, Highest-impact findings, PC/Quest differences, Recommended order, and Approval-required follow-ups.
Keep all reported counts, ranks, dependencies, and blockers tied to the tool result. Do not run an apply-request tool, modify the Unity project, or present an advisory recommendation as an already completed change.
